Uncertainty analysis for an equilibrium partitioning-based estimator of polynuclear aromatic hydrocarbon bioaccumulation potential in sediments

In regulatory evaluations of contaminated sediments, an equilibrium partitioning-based screening test called theoretical bioaccumulation potential (TBP) is often performed to estimate the probable concentrations of neutral organic contaminants that would eventually accumulate in aquatic organisms from continuous exposure to a sediment. The TBP is calculated from contaminant concentration and organic carbon content of the sediment, lipid content of target organisms, and a partition coefficient, usually the biota-sediment accumulation factor (BSAF). However, routine applications of TBP have not included analysis of uncertainty. This paper demonstrates two methods for uncertainty analysis of TBP: a computational method that incorporates random and systematic error and a simulation method using bootstrap resampling of replicated model input parameters to calculate statistical uncertainty measures. For prediction of polynuclear aromatic hydrocarbon (PAH) bioaccumulation in bivalves exposed to contaminated sediments, uncertainty as a factor of TBP ranged from 1.2 to 4.8 using the computational method and 0.5 to 1.9 based on bootstrap 95% confidence intervals. Sensitivity analysis indicated that BSAF parameters, especially tissue contaminant concentration and lipid content, contributed most to TBP uncertainty. In bootstrap tests of significance, TBP significantly over- or underestimated actual PAH bioaccumulation in bivalves in 41% and 10% of comparisons, respectively.